JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 381 Microwave Boosted Glow Discharge Source Using a Slab-line Cavity* Michael Outred Mark H. Rummeli and Edward B. M. Steers SECEAP University of North London Holloway Road London UK N78DB The microwave boosted glow discharge source (GDS) developed by Leis et a/. gives improved analytical performance compared with a Grimm GDS but is somewhat inflexible and unsuitable for more detailed studies of the effects produced by the supplementary discharge. A simple and flexible form of microwave boosted GDS employing a slab-line microwave cavity is described. The source was designed for investigations on fundamental excitation processes in unboosted and microwave boosted GDS. The preliminary results show that the electrical and spectral properties of this source exhibit similar general trends to those of the ‘Leis’ source but the more flexible structure will allow more detailed studies of these properties.Keywords Glow discharge source; microwave boosted discharge; slab-line cavity; charge exchange Direct current glow discharge sources (GDS) offer advantages for bulk and surface analysis of metallic samples but in many cases improved limits of detection are desirable. In a recent review Leis and Steers’ have surveyed the various forms of supplementary discharge that have been used to give significant improvement in detection limits. One of the most convenient methods is that used by Leis et u E . ~ in which a supplementary microwave discharge is generated in a Beenakker coaxial TM, mode cavity3 which forms an integral part of a GDS based on the Grimm s o ~ r c e .~ ~ The intensities of sample resonance lines are greatly increased even though sputter rates are reduced; the background radiation increases by a smaller factor so that improvements in limits of detection by a factor of about 10 can be achieved. In all cases the changes in the spectral output are very complex; they can conveniently be described by the enhance- ment factors F of the various lines where F is defined as the ratio of the intensity with the supplementary discharge to that without the supplementary discharge with constant current and pressure. Table 1 shows typical values for F.’ Thus in addition to the analytical applications of the boosted GDS a study of the spectral changes can aid the study of the excitation processes occurring in the discharge.Table 1 General effects of the supplementary microwave discharge on various groups of lines. (After Leis and Steers’) Typical Emitting species Group of lines values of F Inert gas atoms All lines Inert gas ions All lines Atoms sputtered from the cathode levels Lines from low lying upper Series of upper levels with same electron configuration nl ionization (above or below) Upper levels close to Lines excited directly by Lines excited by cascade from CT excited levels Lines not excited directly or indirectly by CT (upper level too high or too low or CT prevented by selection rules) Ions of cathode material charge transfer (CT) 2-5 0.2-0.5 20-50 Falls with increasing n 100-200 0.1 0.4 3 The structure of the ‘Leis’ source utilizing the Beenakker cavity is somewhat inflexible and the dimensional parameters cannot easily be changed.Moreover side-on observation is difficult. The use of a slab-line cavity6 provides a much more flexible system for studies on discharge processes; various forms of simple d.c. discharge tubes have been constructed located in the slab-line cavity. One typical source (GDS2) is shown in Fig. l(u); in this case a uniform piece of fused silica tubing was used so that the tube could be moved relative to the electrodes and could easily be cleaned or replaced when it became coated with sputtered material. The cavity could be placed at various points along the tube [i.e. the distance D in Fig. l(a) could be varied]. Various cathode sizes have been used with fused silica protective shields to avoid stray dis- charges where necessary so that the discharge always occurs at the end face of the cathode rod. Plane or hollow cathodes can be used.The open structure of the source allows side-on observations to be made though these may be limited by sputtered deposits on the walls. Side-on absorption measure- ments can be carried out in addition to axial measurements made using electrodes with axial holes. There is no significant stray microwave radiation and the plasma impedance can be determined from microwave meas~rernents.~ In an earlier version (GDS1) a shouldered tube [Fig. l(b)] was used with (a) To microwave generator t To vacuum system INlI - Coaxial line Window I 1’ u/O-‘ring Cathbde Water cooling v Anode Parallel rectangular Fused silica tube side plates ( b ) f Cathode face Fig.1 Example of a microwave boosted GDS (GDS2) using the slab-line cavity (see text for details); and (b) sketch of ‘shouldered’ * Presented at the XXVIII Colloquium Spectroscopicum Internationale (CSI) Post-Symposium on Analytical Applications of Glow Discharges in Optical and Mass Spectrometry York UK July 4-7 1993. discharge tube382 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 Microtron 3 - different end blocks but this was more difficult to prepare and proved to be less flexible in use. In this paper preliminary measurements are reported which show that the supplementary microwave discharge affects the electrical and spectral characteristics in the same way as in the Leis source particularly when the latter is used with a ceramic restrictor tube.In the slab-line boosted GDS the microwave discharge is located between the anode and cathode of the d.c. discharge. By contrast in the normal Leis source the micro- wave discharge takes place on the opposite side of the anode tube to the cathode (Fig. 2); if the anode tube is replaced by a ceramic restrictor tube then the surface indicated in Fig. 2 acts as the anode and the microwave discharge lies between the anode and cathode. In the former case the d.c. potential falls with increasing microwave power until a voltage plateau is reached and the intensities also reach a stable value; with the ceramic restrictor tube the voltage can fall to a very low value at high microwave powers so that eventually the lines of the cathode material fall in intensity as the number of sputtered atoms is drastically reduced.This type of behaviour is observed with the slab-line boosted GDS described here. T ~ ~ ~ - T - 20 dB directional- 20 dB directional Slab-line coupler _L c o u p i e r L - cavity Experimental An EMS Microtron 3 was used as the microwave power generator; it has meters that indicate forward and reflected microwave power but the values given are only approximate. The forward power reading is derived from the magnetron current but the efficiency of the magnetron is dependent on the magnitude and phase of the reflected power; the reflected power measurement relies on an uncalibrated crystal detector. The arrangement shown in Fig.3 was therefore used to allow accurate measurements of the net microwave power with tuning stubs to reduce the power reflected back to the Microtron 3. Further details of the commercial equipment used is given in Table 2. The source is connected to a vacuum system with a diffusion pump and liquid nitrogen cooled traps so that the source can Anodehestrictor tube ____._________..__-_.__________.__.______ Anode potential Cathode potential Fig.2 Details of the location of the anode/restrictor tube and elec- trode surfaces in the ‘Leis’ microwave boosted GDS. Using a ceramic restrictor tube the surface marked with a dotted line (and areas further left) acts as the anode Table 2 Equipment used for experimental work ~ ~~ ~~ ~ ~~ ~~ ~- D.c. power supply Microwave power supply Microwave power measurement KSM HVI 2200 (constant current with 10 kQ ballast resistor) E.M.S.Microtron 3 (2.45 GHz 200 W maximum stabilized) MI-Sanders TFT power heads Type 6463 used with 20 dB directional couplers and Type 6460 power meters Kodak Plus X film or E720 photoelectric scanning attachment Spectrographlspectrometer Hilger Medium Quartz with be evacuated to hPa and filled with high-purity neon or argon. The pressure of the filling gas is measured with an MKS Baratron gauge. Plane and hollow cathodes of a h - minium and copper have been employed. Photographic recording of the spectrum was used for overall surveys of the changes produced by the microwave discharge. The plateholder of the medium quartz spectrograph was replaced by the E720 scanning attachment for intensity measurements on individual lines.Results and Discussion The overall effect of the supplementary discharge is apparent in Fig. 4 which shows medium quartz spectrograms for an aluminium cathode in GDS2 with neon as the carrier gas; the same exposure time was used for boosted and unboosted conditions and for the microwave discharge on its own. A number of typical lines have been marked; lines from low- lying A1 I levels (e.g. 257.5 256.7 265.2 266.0 308.2 and 309.3 nm) show a great increase in intensity with the supple- mentary discharge. Those A1 I1 lines which are directly excited by charge transfer (e.g. 263.2nm) fall greatly in intensity whilst the intensity of the 281.6 nm line falls by a smaller amount. The upper level of this line can be populated by a cascade process from charge transfer excited levels and the fact that its intensity falls by a smaller amount than that of lines excited directly by charge transfer suggests that other excitation processes (most probably direct electronic collisional excitation) are also significant for this line.The overall d.c. voltage between anode and cathode affects the energy of the ions and atoms bombarding the cathode (and hence the amount of sputtering) and also the energy acquired by charged particles between collisions (and hence the spectrum). The changes in the overall voltage caused by the microwave discharge are therefore very significant. The variation of enhancement factor with power depends greatly on the line studied. In general lower microwave powers were needed for the smaller diameter discharge tube (GDSl) but the general form of the variations for a particular line were similar in both discharge tubes.Some examples of the variation with microwave power of the voltage and the enhancement factor (derived from photoelectric intensity measurements) for individual lines are given in Figs. 5-8; more detailed results and discussions will be published later. Power Power 1 head I I head I Fig. 3 Schematic arrangement of the microwave measuring equipmentJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 :<- -*. 'X x\ 383 I A1 I1 281.6 Fig. 4 Comparison of d.c. boosted and microwave only discharges GDS2 plane aluminium cathode 10 mm diameter; neon pressure 1.7 hPa; current 8 mA (d.c.and boosted discharges); voltage d.c. 1040 V boosted 610 V net microwave power 7 W (boosted and microwave discharges); exposure time 3 min The exact form of the variations with microwave power depends on the form of GDS used the position of the cavity relative to the discharge (Le. distance D for GDS2) the cathode geometry the filling gas its purity and pressure. In some cases there is a significant plateau region at some point on the voltage versus microwave power characteristic though further power increases cause a marked fall in voltage; in other cases this plateau region is not observed. With hollow cathodes where the production of charge carriers takes place mainly within the cathode the microwave discharge has less effect on the electrical characteristics than for a plane cathode except at low pressures.The form of the intensity versus microwave power character- istic depends to some extent on the voltage uersus power I/ I 0 10 20 30 40 50 Net microwave power/W 1200 1100 3 8 IJ ro 1000 900 800 Fig. 5 Variation of enhancement factor (Cu I 296.1 nm) and voltage with microwave power for the GDS2 plane copper cathode 16 mm diameter; neon pressure 2.0 hPa; current 10 mA - 1 7 0 0 & 0.2 0.1 c I I I I 600 0 10 20 30 40 Net microwave power/W Fig. 6 Variation of enhancement factor (Cu I1 252.7 nm) and voltage with microwave power for GDS2 plane copper cathode 16mm diameter; neon pressure 2.7 hPa; current 16 mA relationship but also on the particular transition studied. Some typical results obtained with a copper cathode in GDS2 with neon as the carrier gas are shown in Figs.5 and 6 together with the variation of discharge voltage with power (cathode diameter 16 mm). For the Cu I 296.1 nm line (Fig. 5 ) there is initially a rapid increase in intensity with net microwave power followed by an extended plateau region (enhancement factor ~ 3 5 ) . On the other hand for the Cu I1 252.7 nm line whose upper level is excited by charge transfer the enhance- ment factor is always less than 1 and falls steadily with increasing microwave power (Fig. 6). Results for the Cu I 327.4nm line show the same trends with GDSl and GDS2 but for lower microwave powers in GDS1. Figs. 7 and 8 show results obtained using an aluminium 0 I 1 2 3 4 5 6 7 Net microwave power/W Fig. 7 Variation of enhancement factor (A1 I 396.2 nm) and voltage with microwave power for GDS1 plane aluminium cathode 10mm diameter; neon pressure 1.3 hPa; current 3 mA 7 300 0 1 2 3 4 5 6 7 Net microwave power/\/\/ Fig.8 Variation of enhancement factor (A1 I 396.2 nm) and voltage with microwave power for GDS1 plane aluminium cathode 10mm diameter; neon pressure 2.6 hPa; current 8 mA384 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 cathode with neon using GDS1. In both cases the intensity of the A1 I 396.2 nm line initially rises with increasing micro- wave power but then falls again as the sputtering rate falls consequent on the reduced overall voltage. At a pressure of 1.3 hPa (Fig. 7) a very high cathode fall is needed to provide enough charge carriers to maintain a current of 3 mA.These charge carriers are readily produced by a very low microwave power so there is initially a sharp fall in the overall voltage. The different forms of the variation of enhancement factor with power in the two cases probably reflect differing contri- butions to the sputtering process from argon ions argon atoms and aluminium ions under the various experimental conditions. The results presented give some examples of the enhance- ment factor versus microwave power relationships and show some effects of the overall voltage. Much more detailed investi- gations are needed; these are in progress and more detailed results including spatial studies will be published in due course. Although the source in its present form is not suitable for analytical applications the flexible structure the ability to apply the microwave field at various positions the potential for making absorption measurements and the wide range of inert gases that can be used in a static system (the source section is sealed during operation) makes it a very useful tool for investigations on discharge processes and for studying the effects of the supplementary discharge. M. H. R. wishes to thank the Science and Engineering Research Council and FI Elemental Winsford Cheshire UK for finan- cial support under an SERC CASE studentship. References 1 Leis F. and Steers E. B. M. Spectrochim Acta Part B 1994 49 289. 2 Leis F. Broekaert J. A. C. and Laqua K. Spectrochim. Acta Part B 1987 42 1169. .3 Beenakker C. I. M. Spectrochim Acta Part B 1976 31 483. 3 Grimm W. Naturwissenschaften 1967 54 586. 5 Grimm W. Spectrochim Acta Part B 1968 23 413. 6 Outred M. and Hammond C. B. Physica Scripta 1976 14 81. 7 Outred M. and Hammond C. B. J. Phys. D. 1980 13 1069. Paper 31050551; Received August 20 1993 Accepted November 5 1993